The question whether the choice of fertilisation procedure may be relevant to reproductive success in poor responder patients is still debated, and clear knowledge of both short- and long-term differences between IVF and ICSI fertilisation is still lacking. Although ICSI was originally indicated for treating couples with severe male factor infertility [13], in recent years some studies suggested that it could provide similar or higher fertilisation rates [14–16] and better embryo quality even in non-male factor couples [17]. Results from our recent study strongly question the superiority of ICSI over conventional IVF in patients with non-male factor infertility and lead us to hypothesise that under specific conditions, ICSI could negatively affect reproductive potential. In 1998, Moreno et al. [9] analyse 96 couples with six or fewer retrieved oocytes and observed that IVF and ICSI in the absence of male infertility factor produce the same results. The same hypothesis was proposed again by other authors [18, 10, 19]. On the other hand, our results are in contrast with Khamsi et al. [17] who observed that ICSI procedure increased fertilisation rate and good-quality embryo rate even in the absence of male factor infertility. In our study we show that on cohorts of poor responder patients with different reproductive age, the use of ICSI decreases reproductive potential in women below 35 years or aged between 35 and 38 years. Although we found no significant differences in fertilisation and cleavage rates in these subgroups of patients, IVF was significantly more advantageous than ICSI for what concerns implantation and PRs. Our results partially agree with those by Fang et al. [20] who retrospectively compared 196 couples undergoing IVF/ICSI cycles with one or two retrieved oocytes with good-prognosis sperm. They found that ICSI patients had higher fertilisation rates although no difference in good-quality embryo rate or PR was noted. The discrepancy between this study and our findings in relation to fertilisation outcome may be ascribed to differences in the size of the cohorts enrolled in the two studies. In fact, our results are consistent with those by Xi et al. (2012) who retrospectively analysed 406 cycles with three or fewer oocytes retrieved from women with similar age undergoing IVF (34.5 ± 4.6 years) or ICSI (36.1 ± 5.5 years) and noted that the PRs and implantation rate were lower in the ICSI group compared with the IVF group [21]. Xi explained the higher PRs and implantation rate after IVF with the lack of ‘natural sperm selection’ when most steps of the fertilisation process are bypassed by sperm injection. In conventional IVF, upon laboratory selection of motile sperm, the sperm which fertilises is further selected through the biological process of sperm–oocyte interaction beginning at the zona pellucida level (ZP) or during sperm penetration through the cumulus matrix [22]. Indeed, the finding that the majority of sperm (average >92 %) bound to the ZP have normal nuclear chromatin DNA strongly suggests that scientist-selected sperm may have a lower quality compared to ZP-bound sperm [23]. Moreover, increased knowledge in the biology of fertilisation process has revealed that sperm–oocyte interaction at the membrane level involves numerous molecular actors with a possible role in sperm fusion and gamete selection [22]. These mechanisms may also act against sperm carrying chromosomal anomaly from paternal origin as well as against chromosomally abnormal oocytes, avoiding the generation of developmentally defective embryos that could lead to pregnancy failure or miscarriage [21, 24]. In addition, decreased chances for successful fertilisation and pregnancy could result from the possible mechanical damage to the oocyte after ICSI [25] as well as the introduction of foreign material such as culture medium or exogenous DNA and infectious material [26]. Finally, our results in reproductively young women can be well explained by considering the relevance of early events of oocyte activation in promoting successful implantation. These include the sperm-induced calcium signal that drives meiosis resumption and embryo development, as well as implantation and postimplantation events, through a specific spatiotemporal pattern of Ca2+ oscillation [27]. Beyond the delivery of sperm DNA, the main reason for ICSI to succeed is that it allows the delivery of PLCz, the sperm component that is capable of generating the fertilisation calcium signal upon sperm fusion [28]. Nevertheless, the finding that abnormal calcium signals were observed in oocytes following ICSI [29, 30] suggests that missing gamete interaction at the surface level in ICSI fertilisation would result in missing or abnormal signalling pathways with a role in subsequent embryonic development. An additional factor with a negative influence on reproductive outcome is ICSI-related risk of parthenogenetic activation caused by oocyte manipulation [31–33]. This may cause asynchrony between the timing of oocyte activation and sperm injection leading to cleavage abnormalities and early embryonic arrest [25]. Our finding that reproductive outcome in patients aged over 38 years undergoing IVF or ICSI was comparable for all parameters analysed strongly indicates that the advantage of IVF over ICSI tends to disappear with the increasing of age. This result can be ascribed to the phenomenon of ovarian ageing responsible for the production of oocytes with a reduced developmental competence related to defective molecular storage, mitochondrial dysfunctions and poor control of chromosome segregation during meiosis [34, 35]. These defects represent a reliable reason why aged oocytes do not benefit from fertilisation mechanisms preserved in IVF and lost in ICSI. Nevertheless, a further reason could be found in the low activation competence of aged oocytes suggested by the observation of abnormal signalling upon exposure to parthenogenetic agents [36].